Research On GNSS/MEMS IMU/Odometer High Precision Fusion Positioning Algorithm In Urban Environment

GNSS system can provide high-precision positioning service for users in outdoor open scene.In the complex urban environment,GNSS signal is seriously disturbed by the environment,so it can't carry out effective positioning,which leads to unreliable positioning accuracy and can't meet the needs of vehicle positioning.The inertial navigation system is less disturbed by the external environment,but the positioning accuracy is limited by the performance of the device itself.When the INS error can't be corrected for a long time,the positioning accuracy is seriously divergent.Aiming at the vehicle application scenario,the odometer information is introduced to increase the redundant observation value of the system and improve the positioning accuracy and reliability of integrated navigation.This paper focuses on the GNSS/MEMS IMU/odometer fusion positioning algorithm in urban environment,including inertial navigation error modeling,low-cost GNSS dual antenna orientation model,vehicle kinematic constraint model and odometer aided fusion positioning algorithm,the main achievements are as follows:(1)The errors of inertial devices directly affect the navigation results.Aiming at the stability problem of inertial devices commonly used in vehicle.The error model of inertial device is established.The system error of inertial device is corrected,and the polynomial fitting model with temperature start-up term error is established.For the random error of inertial devices,Allan variance analysis is used to evaluate the random error parameters of inertial devices.Through the error modeling of inertial device,the device error can be identified and applied effectively.(2)Because the MEMS IMU can't complete the heading initialization when it is static,the subsequent normal navigation will be affected.By establishing a low-cost GNSS dual antenna orientation model,the heading information at the initial time is obtained to assist the MEMS IMU to complete the static initialization.The experimental results show that when the baseline length is 1 m,the precision of low-cost GNSS is better than 0.15 ° and can meet the initialization requirements of MEMS IMU.(3)Through the reasonable use of vehicle motion characteristics and state,the positioning accuracy in complex environment can be improved.The nonholonomic constraint model is established by using the speed direction characteristics of the vehicle,and the Zero-velocity constraint model is established by using the Zero-velocity characteristics of the vehicle.The results show that in the open environment,the positioning accuracy is improved by using the vehicle kinematic model constraints.In the shadowing environment,the nonholonomic constraint model can effectively restrain the divergence of inertial navigation devices and improve the positioning accuracy.(4)For vehicle applications,the introduction of odometer information aided fusion positioning can greatly improve the positioning performance of INS.The positioning model based on GNSS/MEMS IMU/odometer fusion is established.The positioning accuracy of GNSS positioning model and fusion positioning model is analyzed by testing in urban environment.The results show that both of them can meet the requirements of high-precision positioning in open environment.In the complex environment,GNSS can't effectively locate,and the fusion positioning accuracy is obviously better than GNSS positioning accuracy,and the fusion positioning accuracy is more reliable,which can meet most of the high-precision positioning requirements.